When attempting to control sound, there are in fact two aspects one must consider: the noise or echo inside the room, and the noise that we are trying to contain from either entering or escaping the room. Both are interdependent and are addressed using a different solution. Excessive reverb or echo in an overly ‘live’ room makes communication difficult. This is particularly acute in boardrooms, classrooms and conference rooms where understanding each word is critical to ensuring the message being transmitted is clearly understood. This is referred as intelligibility. A problem occurs when the direct sound from the orator’s voice is interfered with echoes reflecting off of hard surfaces such as windows, walls, ceilings and table tops. The following image shows the direct sound (green) and the reflected sound (red) coming off a hard surface that will arrive a few milliseconds later. When the intensity of the echo approaches the amplitude of the direct sound, the two sounds combine, introducing ‘peaks and valleys’ known as comb-filtering whereby various frequencies amplify each other when in phase, or cancel each other out when out of phase.  

ceiling-reflect
energy-of-reflections2

This image takes a more detailed look at the phenomenon whereby the initial sound (green) arrives first, then a series of powerful first order reflections from the walls, windows and ceiling arrive. This is followed by a trailing series of primary reflections that create a reverberant field. When the room reverb exceeds one second, voice localization becomes difficult, reducing our ability to comprehend what is being said.

The following graph shows the relative energy of a typical male voice at two different amplitudes. The blue line depicts a normal speaking voice while the red line depicts a raised voice. As the voice level increases, the energy tends to gather in the mid range, between 300Hz and 1500Hz.

voice-energy

To further illustrate the problem, one merely needs to superimpose the frequency range of the human voice on top of the absorption coefficient of glass and gypsum board. The graph below does this. These hard surfaces absorb less than 5% in the voice range, resulting in almost 100% of the energy being reflected back into the room.

compare-voice-glass

The usual approach to solving the problem is mounting absorptive acoustic panels to the walls. But sometimes, treating the walls may be impractical. For instance, a classroom may have windows and blackboards that obviously cannot be covered while a boardroom may be surrounded by windows, bookcases and wall ornaments. In these cases, limited space often makes treating the ceiling the only option. Ideally, one would employ a mixture of both.